Issue link: https://iconnect007.uberflip.com/i/455818
February 2015 • SMT Magazine 31 eLeCTrOSTaTIC MeCHaNISM OF NuCLeaTION aNd GrOWTH OF MeTaL WHISKerS continues Feature • Growth can be promoted by thermal cy- cling. • Growth rate is zero below a threshold film thickness and approaches zero for high film thickness. It appears to be zero for bulk tin. • For sputtered films, the growth rate ap- pears to be a minimum for near-zero resid- ual stress, and greater for tensile as well as compressive stress. • Growth rate is somewhat higher at high humidity. • Growth rate seems to be higher from fine- grained microstructure. • Growth rate can be increased by some kinds of residues on the surface. • Most metals dissolved in tin appear to in- crease growth rate. The one exception is Pb. The mechanism may have to do with altering the grain structure to equiaxed (from columnar). • I do not recall hearing of the effect of small amounts of Pb (~1%) in Sn for vapor-de- posited films, or even for very thin electro- plated films. • Distribution of thickness and length are log-normal. • There appears to be no correlation between thickness and length. • Median thickness is about 3 μm. • Longest whisker reported: ~25 mm. • Thinnest and thickest whiskers reported: ~100 nm, ~20 μm. • Various growth morphologies: needle-like, odd-shaped eruptions, occasional branch- es, and there may longitudinal or circum- ferential striations. Acicular (needle-like) whiskers may be bent or kinked, and may not have the same thickness along the en- tire length. • Long whiskers are in constant motion in air—can be compared to Brownian mo- tion. • Whiskers have an oxide coating ~1–3 nm thick, even in vacuum. (Growth rate is log- arithmic.) • A whisker that melts exits the skin, leaving it behind. • Whiskers penetrate even a thick oxide film (grown by prolonged exposure to steam). • Whiskers eventually penetrate polymer (including Parylene) coatings, with the ap- parent exception of "whisker-tough." • Whiskers appear to not penetrate thin caps of certain metals, and readily penetrate thicker caps of other metals. • Whiskers appear to not penetrate thin films of tailored ceramics produced by chemical vapor deposition if the substrate has been properly prepped. To emphasize the most challenging ques- tions, here is the author's short list: • A mystery of high aspect ratios, height/ diameter up to ~10,000 not seen in other physics. Why wouldn't metal whiskers col- lapse into spheres, as other droplets do to minimize surface energy? • Is their relation to metals of essence? In other words, why are metal whiskers met- al? • What is behind the metal whiskers ran- domness? Why do they grow here but not there, why are their parameters so dis- persed, and what makes it so difficult to controllably grow or predict their appear- ance? • What does Pb do in suppressing whiskers? Multiple attempts to understand the mecha- nisms of whiskers growth revolved around the role of surface stresses relived by whisker pro- duction, dislocation effects, oxygen reactions, and recrystallization. It was shown stress gradi- ents along with certain assumptions about sys- tem parameters can explain tin whisker growth rates but not their existence, shapes and sta- tistics. Overall, these attempts have not led to verifiable quantitative predictions. The 60-year old whisker challenge thus re- mains outstanding against the background of other historical developments in natural sci- ences. As an example, the fundamentally new phenomenon of superconductivity was discov- ered in 1911 and explained in 1957, taking a shorter time to understand than metal whiskers, in spite of being the first encounter of the mac- roscopic quantum phenomena physics. This remarkable elusiveness of the metal whisker